Natural Language Processing

1
Natural Language Processing

• (Ling 581)

2
What is NLP?

• Natural language processing
• Computer methods and their usefulness (or
  uselessness) for human language processing
  (textual, spoken, gestural, etc.)
• Implementation of techniques, procedures,
  algorithms for language computation
• AI and cognitive science w/rt human communication
• Enabling human-machine communication
• Enhancing human-human communication

3
NOT this NLP!

• Neuro-Linguistic Programming
• New-age, pop-psychology trance-inducing hypnosis
  technique for behavior modifi-cation
• Improving sales techniques
• Enhancing sports performance
• Losing weight
• Dealing with problem people

4
Situating NLP
philosophy
linguistics
computer science
NLP
communication
math/statistics
psychology/cognitive science
5
Theoretical foundations

• math statistics, calculus, algebra, modeling
• computational paradigms connectionist,
  rule-based, cognitively plausible
• linguistics LFG, HPSG, GB, OT, CG, etc.
• architectures stacks, automata, networks,
  compilers

6
Multidisciplinarity

• humanities, natural and behavioral sciences, and
  engineering
• linguistics, computer science, psychology, and
  mathematics
• theory and practice, science and art
• models, foundations vs. corpora, data (top-down
  vs. bottom-up)

7
Some areas of research

• Corpora, tools, resources, standards
• Language/grammar engineering
• Machine (assisted) translation, tools
• Language modeling
• Lexicography
• Speech

8
NLP systems

• Computer systems designed to perform 1 language
  functions (perhaps with other tasks)
• LISP, Prolog, C, C, Perl, Java, specialized
  high-level metalanguages
• Downloaded, web-enabled, proprietary
• Various UIs, resource requirements, complexity
  of use
• Mono/multilingual, various i/o modalities

9
Sample NLP class tasks

• Computational morphology finite-state techniques
  for morphophonological tasks
• Lexical ontologies using WordNet
• Language modeling neural net, memory-based
  learning, or AML
• Language generation generate full sentences from
  semantic frames
• Parsing using finite-state parsers and
  transducers for syntax and semantics
• Speech recognition spoken language
  identification
• Speech synthesis develop a TTS system for a
  language